P1: GXB

Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512 November 22, 2003 15:36 Style file version June 25th, 2002

Journal of Psychopathology and Behavioral Assessment, Vol. 26, No. 1, March 2004 (
C 2004)

Multidimensional Assessment of Emotion Regulation and

Dysregulation: Development, Factor Structure, and Initial
Validation of the Difficulties in Emotion Regulation Scale1

Kim L. Gratz2,3,4 and Lizabeth Roemer2

Accepted August 8, 2003

Given recent attention to emotion regulation as a potentially unifying function of diverse symptom
presentations, there is a need for comprehensive measures that adequately assess difficulties in emo-
tion regulation among adults. This paper (a) proposes an integrative conceptualization of emotion
regulation as involving not just the modulation of emotional arousal, but also the awareness, under-
standing, and acceptance of emotions, and the ability to act in desired ways regardless of emotional
state; and (b) begins to explore the factor structure and psychometric properties of a new measure, the
Difficulties in Emotion Regulation Scale (DERS). Two samples of undergraduate students completed
questionnaire packets. Preliminary findings suggest that the DERS has high internal consistency, good
testretest reliability, and adequate construct and predictive validity.

KEY WORDS: emotion regulation; emotion dysregulation; assessment; experiential avoidance; deliberate
self-harm.

Consistent with behavioral theories of psychopathol-
ogy that highlight the importance of the function of
problem behaviors rather than symptom picture (Hayes,
Wilson, Gifford, Follette, & Strosahl, 1996), increasing
attention has been paid to emotion regulation as a poten-
tially unifying function of diverse symptom presentations
and maladaptive behaviors (Gross & Munoz, 1995). Al-
though there is some preliminary literature on the role
of emotion regulation deficits in a range of clinical dis-
orders, including substance abuse (Hayes et al., 1996),
generalized anxiety disorder (Mennin, Heimberg, Turk,
& Fresco, 2002), and complex posttraumatic stress disor-
der (Cloitre, 1998), the most comprehensive work high-
lighting the role of emotion dysregulation in a clinical
1 This
research was part of the first authors dissertation, and portions
were previously presented at the annual meeting of the Association for
Advancement of Behavior Therapy in November, 2001.
2 Department of Psychology, University of Massachusetts Boston,
Boston, Massachusetts.
3 Present address: Center for the Treatment of Borderline Personal-
ity Disorder, McLean Hospital, Harvard Medical School, Boston,
Massachusetts.
4 To whom correspondence should be addressed at McLean Hospital, 115
Mill Street, Belmont, Massachusetts 02478; e-mail: klgratz@aol.com.
disorder has been Linehans theoretical work (Linehan,
1993) on the development of borderline personality disor-
der. Linehan proposes that emotion dysregulation is one
of the central features of borderline personality disorder
and underlies many of the associated behaviors of this dis-
order, including deliberate self-harm (a behavior thought
to serve an emotion-regulating function). Her conceptu-
alization of self-harm as an emotion regulation strategy
is supported by both empirical and theoretical literature
on the function of this behavior (see Briere & Gil, 1998;
Gratz, 2003). It has been similarly suggested that the per-
petration of violence toward others (e.g., intimate part-
ners) may function to regulate emotions (Jakupcak, Lisak,
& Roemer, 2002)consistent with findings from experi-
mental studies that aggressive behavior can serve an affect
regulatory function (Bushman, Baumeister, & Phillips,
2001).
Despite its clinical significance, however, the role of
emotion regulation deficits in the development and main-
tenance of clinical difficulties has not been adequately
researched in adults. Likely contributing to the lack of
research in this area is the absence of both consistent,
agreed-upon conceptualizations of emotion regulation
and comprehensive measures that adequately assess the

41
0882-2689/04/0300-0041/0
C 2004 Plenum Publishing Corporation
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512 November 22, 2003
42
complexity of this construct among adults. The purpose
of the present study was to develop and validate a mea-
sure of clinically relevant difficulties in emotion regula-
tion that is based on a comprehensive, integrative concep-
tualization of emotion regulation. This paper (a) reviews
the extant literature on the conceptualization and mea-
surement of emotion regulation (and dysregulation); (b)
provides an integrative conceptualization of emotion reg-
ulation that may be used in future research in this area; and
(c) begins to explore the factor structure and psychomet-
ric properties of a new measure of difficulties in emotion
regulation, the Difficulties in Emotion Regulation Scale
(DERS).
Some conceptualizations of emotion regulation em-
phasize the control of emotional experience and expres-
sion (especially the expressive control of negative emo-
tions) and the reduction of emotional arousal (see Cortez
& Bugental, 1994; Garner & Spears, 2000; Kopp, 1989;
Zeman & Garber, 1996). In contrast, others emphasize
the functional nature of emotions in conceptualizing emo-
tion regulation, suggesting that emotion regulation is
not synonymous with emotional control and, as such,
does not necessarily involve immediately diminishing
negative affect (Cole, Michel, & Teti, 1994; Thompson,
1994). These latter approaches suggest that deficien-
cies in the capacity to experience (and differentiate)
the full range of emotions and respond spontaneously
may be just as maladaptive as deficiencies in the abil-
ity to attenuate and modulate strong negative emotions
(Cole et al., 1994; Gross & Munoz, 1995; Paivio &
Greenberg, 1998). Similarly, some researchers have sug-
gested that adaptive emotion regulation involves moni-
toring and evaluating emotional experience in addition to
modifying it, highlighting the importance of the awareness
and understanding of emotions (Thompson & Calkins,
1996).
The emphasis on the functionality of emotions is
consistent with theory and research highlighting the po-
tentially paradoxical, dysregulating effects of attempts to
control emotional experience and expression (despite the
fact that such emotional control has often been equated
with emotion regulation, e.g., Garner & Spears, 2000;
Zeman & Garber, 1996). Hayes et al. (1996) have sug-
gested that efforts to avoid internal experiences (e.g., un-
wanted thoughts and feelings) underlie many psycholog-
ical disordersa theory with growing empirical support
(see Stewart, Zvolensky, & Eifert, 2002). Further, both the
general tendency to constrict emotional expression and ex-
perimental instructions to conceal ones emotional expres-
sions have been associated with increased physiological
arousal (Notarius & Levenson, 1979; Gross & Levenson,
1997), suggesting that attempts to control emotional ex-
pression may increase risk for emotion dysregulation
15:36 Style file version June 25th, 2002
Gratz and Roemer
(given that high levels of arousal are more difficult to reg-
ulate; see Eisenberg, Cumberland, & Spinrad, 1998; Flett,
Blankstein, & Obertynski, 1996). This literature suggests
that an emphasis on the control, rather than the accep-
tance, of emotional responses may confound processes
that undermine regulation with those that are regulatory.
Therefore, some conceptualizations of emotion regulation
emphasize the importance of accepting and valuing emo-
tional responses (Cole et al., 1994; Linehan, 1993). Con-
sistent with these approaches, researchers have suggested
that the tendency to experience negative emotions in re-
sponse to ones own emotional reactions (indicative of a
lack of emotional acceptance) is maladaptive, and associ-
ated with greater difficulties in emotion regulation (Cole
et al., 1994; Hayes, Strosahl, & Wilson, 1999; Paivio &
Greenberg, 1998).
Researchers have suggested the necessity of consid-
ering the demands of the situation and goals of the indi-
vidual when evaluating emotion regulation (Thompson,
1994; Thompson & Calkins, 1996), as emotion regulation
can only be understood and evaluated within a specific
context (Cole et al., 1994; Thompson, 1994). Knowledge
of the specific emotion regulation strategies used by an
individual, in the absence of information on the context in
which they are used, may provide little information about
the individuals ability to regulate her or his emotions ef-
fectively. Adaptive emotion regulation instead involves
flexibility in the use of emotion regulation strategies (Cole
et al., 1994; Thompson, 1994).
Researchers have also suggested that adaptive emo-
tion regulation involves altering the intensity or duration of
an emotion rather than changing the discrete emotion that
is experienced (Thompson, 1994; Thompson & Calkins,
1996). In other words, adaptive regulation involves modu-
lating the experience of emotions rather than eliminating
certain emotions. This modulation of arousal is thought to
be in the service of reducing the urgency associated with
the emotion so that the individual is able to control her
or his behavior (as opposed to controlling emotions them-
selves). These conceptualizations of emotion regulation
emphasize the ability to inhibit inappropriate or impulsive
behaviors, and behave in accordance with desired goals,
when experiencing negative emotions (see Linehan, 1993;
Melnick & Hinshaw, 2000).
On the basis of the above conceptual and empir-
ical work, emotion regulation may be conceptualized
as involving the (a) awareness and understanding of
emotions, (b) acceptance of emotions, (c) ability to
control impulsive behaviors and behave in accordance
with desired goals when experiencing negative emotions,
and (d) ability to use situationally appropriate emotion
regulation strategies flexibly to modulate emotional
responses as desired in order to meet individual goals and
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512
Assessment of Difficulties in Emotion Regulation
situational demands.5 The relative absence of any or all of
these abilities would indicate the presence of difficulties
in emotion regulation, or emotion dysregulation.
Despite the clinical utility of assessing difficulties in
emotion regulation, there are few measures of emotion
regulation or dysregulation in adults. Most of the mea-
sures used to assess emotion regulation deficits are actu-
ally measures of closely-related constructs, leading some
researchers to use multiple measures in an attempt to cap-
ture the construct of emotion regulation (see, e.g., Mennin
et al., 2002).
The most commonly used measure of emotion regu-
lation is Catanzaro and Mearns Generalized Expectancy
for Negative Mood Regulation Scale (NMR; Catanzaro
& Mearns, 1990). The NMR measures beliefs that some
behavior or cognition will alleviate a negative state or
induce a positive one (p. 547), emphasizing the elimi-
nation and avoidance of negative emotions (as opposed
to the ability to act in desired ways in the presence
of these emotions). Many items on the NMR refer to
particular emotion regulation strategies that people may
use to modulate emotional arousal. This practice implies
that emotion regulation strategies are adaptive regardless
of context and that certain strategies are more adaptive
than others universallyan implication that is inconsis-
tent with conceptualizations of emotion regulation that
emphasize the contextually dependent nature of adaptive
regulation strategies (see Cole et al., 1994; Thompson,
1994; Thompson & Calkins, 1996). Moreover, many of
the strategies listed in the NMR seem to equate emotion
regulation with emotional avoidance (e.g., When Im up-
set, I believe that I wont be able to put it out of my mind
(negatively scored); or When Im upset, I believe that I
can forget about whats upsetting me pretty easily). Al-
though the NMR does not assess some potentially impor-
tant aspects of emotion regulation (such as the awareness,
understanding, and acceptance of emotions), it does mea-
sure one important aspect of emotion regulation as con-
ceptualized here: access to emotion regulation strategies
perceived as effective.
Another measure occasionally used to capture the
construct of emotion regulation is the Trait Meta-Mood
Scale (TMMS; Salovey, Mayer, Goldman, Turvey, &
Palfai, 1995), which assesses individual differences in
the ability to reflect upon and manage ones emotions
5 Inaddition to its relationship to several conceptualizations of emotion
regulation reviewed above, this definition shares many characteristics
with Saarnis definition of emotional competence, which includes
awareness of ones emotional experience, the capacity for adaptive cop-
ing in response to distressing contexts or when experiencing aversive
emotions, and acceptance and allowance of ones emotional experience
(Saarni, 1999).
November 22, 2003 15:36 Style file version June 25th, 2002
43
(Salovey et al., 1995, p. 126). Participants rate the degree
to which they endorse a series of statements regarding
their attitudes toward, and experience of, their emotions.
Factor analyses have yielded three subscales (attention to
feelings, clarity of feelings, and mood repair), each of
which demonstrates good internal consistency and con-
vergent validity. The TMMS assesses some aspects of
emotion regulation that are absent from the NMR (e.g.,
emotional awareness and understanding), whereas the re-
pair subscale is similar to the NMR in its emphasis on
emotional avoidance (e.g., I try to think good thoughts
no matter how badly I feel). The TMMS also does not
include an assessment of the ability to engage in desired
behaviors when experiencing negative emotionsthe ab-
sence of which may be particularly important in the clin-
ical realm.
The Difficulties in Emotion Regulation Scale
(DERS) was developed to assess emotion dysregulation
more comprehensively than existing measures. The DERS
items were chosen to reflect difficulties within the follow-
ing dimensions of emotion regulation: (a) awareness and
understanding of emotions; (b) acceptance of emotions;
(c) the ability to engage in goal-directed behavior, and
refrain from impulsive behavior, when experiencing neg-
ative emotions; and (d) access to emotion regulation strate-
gies perceived as effective. The final dimension reflects an
attempt to measure the flexible use of situationally appro-
priate strategies to modulate emotional responses. Given
the aforementioned limitations associated with the prac-
tice of delineating specific regulation strategies when at-
tempting to measure emotion regulation (in particular, the
implication that certain strategies are more adaptive than
others regardless of context), the subjective appraisal of
effectiveness was assessed instead, with the hope that this
would take into account the contextually dependent nature
of adaptive regulation strategies. The primary purpose of
the two studies reported below was to begin to explore the
factor structure and psychometric properties of the DERS,
thereby aiding in the development of a theoretically-based,
comprehensive, empirically validated measure of emotion
dysregulation. The first study examined the factor struc-
ture, internal consistency, and construct and predictive va-
lidity of the DERS, and a second study examined the mea-
sures testretest reliability.
STUDY 1
Method
Participants
Questionnaire packets were distributed to 479 stu-
dents from undergraduate psychology courses offered at
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512 November 22, 2003
44
the University of Massachusetts Boston. Of these, 373
packets were returned, resulting in a response rate of 78%.
There were no significant differences in age, racial back-
ground, or gender between participants who returned the
packets and those who did not. As they were missing data
on one or more of the variables of interest, 16 participants
were excluded from the following analyses.
The final sample of 357 participants ranged in age
from 18 to 55 years, with a mean age of 23.10 (SD =
5.67). Seventy-three percent (n = 260) of these partici-
pants were female. Sixty-five percent of participants self-
identified as White, 17% as Asian, 8% as Black/African
American, 4% as Hispanic, and 6% reported another or
unspecified racial background. Participants were predomi-
nantly single (89%) and heterosexual (90%). There was lit-
tle difference, demographically, between participants who
completed all of the measures and those who did not.
Measures
Difficulties in Emotion Regulation Scale. The initial
Difficulties in Emotion Regulation Scale (DERS) was a
41-item, self-report measure developed to assess clini-
cally relevant difficulties in emotion regulation. The DERS
items were developed and selected on the basis of numer-
ous conversations with colleagues well versed in the emo-
tion regulation literature. The NMR (Catanzaro & Mearns,
1990; see later) was used as a template and helped to struc-
ture the format of some of the items (although not the
content of the items). Specifically, in order to assess diffi-
culties regulating emotions during times of distress (when
regulation strategies are most needed), many items begin
with When Im upset, similar to the NMR. The content
of the items differ, however, as DERS items were chosen
to reflect difficulties within the following dimensions of
emotion regulation: (a) awareness and understanding of
emotions; (b) acceptance of emotions; (c) the ability to
engage in goal-directed behavior, and refrain from impul-
sive behavior, when experiencing negative emotions; and
(d) access to emotion regulation strategies perceived as
effective. Participants are asked to indicate how often the
items apply to themselves, with responses ranging from 1
to 5, where 1 is almost never (010%), 2 is sometimes (11
35%), 3 is about half the time (3665%), 4 is most of the
time (6690%), and 5 is almost always (91100%). For the
purposes of this study, DERS items were recoded so that
higher scores in every case indicated greater difficulties in
emotion regulation (i.e., greater emotion dysregulation).
Generalized Expectancy for Negative Mood Regu-
lation Scale. The Generalized Expectancy for Negative
Mood Regulation Scale (NMR; Catanzaro & Mearns,
15:36 Style file version June 25th, 2002
Gratz and Roemer
1990) is a 30-item, self-report questionnaire used to
measure expectancies for the self-regulation of negative
moods. The NMR asks participants to indicate the extent
to which they believe that their attempts to alter their neg-
ative moods will work. The NMR has high internal con-
sistency, and adequate construct and discriminant validity
(Catanzaro & Mearns, 1990). The NMR also has adequate
testretest reliability over periods of 34 weeks (r = .74
for women; r = .76 for men) and 68 weeks (r = .78 for
women; r = .67 for men; Catanzaro & Mearns, 1990).
Items were recoded so that higher scores in every
case indicated higher expectancies for negative mood reg-
ulation (or greater emotion regulation), and a sum was cal-
culated. Given that the NMR is the most commonly used,
empirically supported measure of emotion regulation, it
was included in the present study to assess the construct
validity of the DERS.
Acceptance and Action Questionnaire. The Accep-
tance and Action Questionnaire (AAQ; Hayes et al.,
in press), is a nine-item, self-report measure of experi-
ential avoidance (i.e., the tendency to avoid unwanted in-
ternal experiences, such as emotions and thoughts). Items
assess experiential avoidance and control, negative eval-
uations of internal experience, psychological acceptance
(or lack thereof), and the extent to which an individual
acts regardless of emotional distress or, conversely, allows
distress to prompt behavioral avoidance (Hayes et al., in
press). Sample items include I rarely worry about getting
my anxieties, worries, and feelings under control (reverse
scored) and If I could magically remove all the painful
experiences Ive had in my life, I would do so. The AAQ
is significantly correlated with the tendency to engage in
the suppression of thoughts in both clinical and nonclinical
populations (r s = .44.50; Hayes et al., in press), provid-
ing preliminary evidence of its construct validity. There is
also evidence that the AAQ assesses a unique, clinically
relevant construct, as it remains significantly correlated
with measures assessing a wide range of psychopathology
when controlling for social desirability and the tendency
to suppress thoughts (Hayes et al., in press).
Items on the AAQ were recoded so that higher scores
in every case indicated greater experiential avoidance, and
a sum was calculated. The AAQ was included to assess
the construct validity of the DERS, and was expected to be
positively associated with DERS scores (see Cole et al.,
1994; Hayes et al., 1999; Hayes et al., 1996; Linehan,
1993; Paivio & Greenberg, 1998).
Emotional Expressivity Scale. The Emotional Ex-
pressivity Scale (EES; Kring, Smith, & Neale, 1994) is
a 17-item, self-report questionnaire that assesses general
emotional expressivity. The EES is based on the concep-
tual definition of emotional expressiveness as the extent
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512
Assessment of Difficulties in Emotion Regulation
to which people outwardly display their emotions (p. 936,
Kring et al., 1994), regardless of the valence of the emo-
tion (i.e., positive or negative) or the manner in which it is
expressed (i.e., facially, vocally, or gesturally). The EES
has high internal consistency, and adequate convergent,
discriminant, and construct validity (Kring et al., 1994).
The EES is significantly correlated with spontaneous emo-
tional expressiveness in the laboratory (as assessed by ob-
server ratings of general expressivity in response to emo-
tionally evocative film clips) among both college students
(r = .38) and community residents (r = .43; Kring et al.,
1994). The EES also has good testretest reliability over
a 4-week period (r = .90; Kring et al., 1994).
Items on the EES were recoded so that higher scores
in every case indicated greater emotional expressivity, and
a sum was calculated. The EES was included to assess the
construct validity of the DERS. Contrary to some defini-
tions of emotion regulation (see e.g., Garner & Spears,
2000; Zeman & Garber, 1996), the conceptualization of
emotion regulation on which the DERS is based does not
equate emotion regulation with expressive control. On the
contrary, emotional expressiveness is expected to facilitate
emotion regulation (see Linehan, 1993). Thus, the DERS
was expected to be negatively correlated with EES scores
of emotional expressivity.
Deliberate Self-Harm Inventory. The Deliberate
Self-Harm Inventory (DSHI; Gratz, 2001) is a behav-
iorally based, self-report questionnaire that assesses de-
liberate self-harm (the deliberate, direct destruction or al-
teration of body tissue without conscious suicidal intent,
but resulting in injury severe enough for tissue damage
to occur). This measure is composed of 17 items and as-
sesses various aspects of deliberate self-harm, including
frequency, severity, duration, and type of self-harming be-
havior. The DSHI has high internal consistency ( = .82),
adequate construct, convergent, and discriminant validity,
and adequate testretest reliability over a period ranging
from 2 to 4 weeks, with a mean of 3.3 weeks ( = .68,
p < .001; Gratz, 2001).
For the present study, a continuous variable was cre-
ated to measure frequency of reported self-harm behavior.
Participants scores on the frequency questions for each
of the 17 items were summed to create a variable of the
total frequency of self-harm behavior. Given the growing
body of literature on the emotion regulating function of
self-harm behavior (see Gratz, 2003; Linehan, 1993), the
DSHI was included in the present study to provide some
preliminary data on the predictive validity of the DERS.
The DERS was expected to be positively associated with
frequency of self-harm behavior.
Abuse-Perpetration Inventory. The Abuse-Perpet-
ration Inventory (API; Lisak, Conklin, Hopper, Miller,
November 22, 2003 15:36 Style file version June 25th, 2002
45
Altschuler, & Smith, 2000) is a self-report, behaviorally
based instrument that assesses childhood physical and
sexual abuse as well as four types of perpetration. The API
has been shown to have adequate criterion and construct
validity (Lisak et al., 2000). The section of the API
assessing perpetration behavior, the Perpetration History
(PH) questionnaire, was used to assess the extent to which
an individual had engaged in intimate partner abuse.
The PH is comprised of 28 items assessing four types
of perpetration, including physical and sexual abuse of
children and adults. Seven of these items describe specific
experiences of intimate partner abuse, ranging from slap-
ping ones partner to attempting to kill ones partner with
a knife or gun. For each question answered yes, partic-
ipants answer several follow-up questions regarding their
age at the time, the age of their victim, and the frequency
of the behavior. For this study, participants scores on the
frequency questions for each of the seven items pertaining
to intimate partner abuse were summed to create a vari-
able of the total frequency of partner abuse. Given recent
suggestions that the abusive behavior of men toward an
intimate partner may function as an attempt to regulate
emotions (e.g., by reducing distressing arousal; Jakupcak
et al., 2002), the PH was included in the present study
to provide preliminary data on the predictive validity of
the DERS. Although the emotion regulating function of
partner abuse has been suggested only for men, this rela-
tionship was examined among women as well.
Procedure
Participants were informed fully, both verbally and
in writing, about the purpose of the study prior to partici-
pation, and signed a consent form describing the purpose
of the study, the potentially distressing subject matter, and
the confidential nature of the data (i.e., a code number was
assigned to each questionnaire before the data were exam-
ined, the data were stored in a secure area, and access to
the data was limited to the research team). Students who
chose to participate in the study completed questionnaire
packets consisting of the measures described above (see
Measures) and other measures unrelated to the current
study. Participants received research credits in exchange
for their participation.
Results
Factor Structure
Exploratory factor analysis was used to provide pre-
liminary data on the factor structure of the DERS and
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512 November 22, 2003 15:36 Style file version June 25th, 2002

46 Gratz and Roemer

identify the underlying dimensions of emotion regula- Table I. Factor Loadings for the 36 DERS Items Included in the Final
tion as assessed by this scale. As the conceptualization Factor Analysis (N = 357)
of emotion regulation on which the DERS was based is
Factor
multifaceted, it was expected that this measure would be
composed of different (albeit associated) factors, each re- Item 1 2 3 4 5 6
flecting a dimension of emotion regulation wherein diffi- DERS29 .91 .09 0.05 .14 .19 .10
culties may occur. DERS25 .76 .07 0.01 .03 .06 .06
Preliminary Analyses. Prior to conducting the factor DERS15 .72 .07 0.07 .03 .07 .15
analyses, the response distributions of all individual DERS DERS14 .53 .06 0.10 .01 .19 .04
DERS33 .48 .04 0.09 .03 .35 .05
items were examined. As none of these variables were
DERS27 .41 .04 0.11 .11 .25 .14
excessively skewed or kurtotic (Kendall & Stuart, 1958), DERS30 .14 .88 0.10 .03 .10 .03
no items were excluded from the analyses on the basis of DERS22 .00 .88 0.05 .16 .09 .11
their response distributions. However, upon inspection of DERS16 .01 .85 0.03 .03 .04 .02
the correlation matrix for all items, item 13 (When Im DERS38 .07 .65 0.06 .10 .18 .05
DERS24 .17 .64 0.11 .25 .04 .01
upset, I allow myself to feel that way) was excluded from
DERS37 .08 .10 1.00 .06 .11 .11
future analyses on the basis of its low correlations with DERS31 .05 .05 0.79 .09 .04 .11
the overall scale score (r = .06, ns), as well as the other DERS17 .03 .05 0.75 .04 .02 .13
scale items (mean r with other items = .01; 36 of 40 r s DERS23 .06 .00 0.58 .14 .24 .18
below .20). The decision to omit this item on the basis DERS4 .09 .04 0.52 .06 .02 .23
DERS28 .10 .10 0.40 .27 .08 .08
of its minimal association with the other variables was
DERS7 .01 .04 0.02 .74 .01 .08
supported by the fact that, in a preliminary factor analysis DERS3 .11 .00 0.04 .67 .04 .21
of all items, the communality of item 13 upon extraction DERS12 .15 .08 0.01 .61 .05 .04
was below .30 (the only item for which this was the case). DERS21 .17 .05 0.03 .59 .13 .07
Factor Analyses. Responses to the remaining DERS9 .00 .14 0.07 .58 .06 .08
DERS39 .06 .15 0.10 .57 .09 .06
40 items were subjected to a common factor analysis (see
DERS20 .01 .08 0.01 .01 .86 .04
Floyd & Widaman, 1995), using the principal axis factor- DERS19 .07 .08 0.04 .06 .79 .13
ing method of extraction and promax oblique rotation to DERS35 .16 .01 0.02 .05 .64 .11
allow for correlations among factors. DERS40 .00 .18 0.02 .07 .61 .09
Given Floyd and Widamans suggestion that the scree DERS32 .17 .03 0.02 .07 .59 .02
DERS26 .09 .21 0.03 .43 .49 .21
test is a more accurate method for retaining factors than DERS41 .03 .20 0.19 .09 .45 .02
the more commonly used criterion of eigenvalues >1.00 DERS34 .34 .06 0.08 .07 .45 .08
(the Kaiser-Guttman criterion), the scree test was used in DERS6 .08 .06 0.02 .01 .10 .81
the present study (Floyd & Widaman, 1995). The scree test DERS5 .05 .07 0.25 .04 .16 .71
suggested retaining six or seven factors, and thus analyses DERS10 .03 .05 0.12 .06 .20 .69
DERS8 .06 .07 0.12 .32 .05 .59
were conducted to examine the interpretability and utility DERS1 .03 .00 0.13 .38 .10 .42
of both six- and seven-factor solutions. As the six-factor
solution was more interpretable (and the seven-factor so-
Note: Items loading on each factor are in boldface.
lution contained a factor on which only two variables had
primary loadings), six factors were retained in the final are things I can do to manage my emotions) had a double
analyses. loading on two factors (with loadings of .46 and .50 on
In regard to item selection for the six factors, factor factors 4 and 6, respectively).
loadings of .40 and higher were considered meaningful. After excluding these four items, the factor analysis
Items with loadings below .40 on all factors were excluded was recomputed on the remaining 36 items to ensure that
from further analyses. In addition, variables that appeared all had factor loadings of .40 or higher (see Table I for
to have double loadings on two factors (i.e., loadings of the final factor loadings).6 Upon extraction, the six factors
greater than .40 and of comparable size) were excluded. accounted for 55.68% of the total variance of the measured
On the basis of these criteria, four items were deleted fol- variables (see Table II for the eigenvalues and percentage
lowing the first factor analysis. Items 2 (I feel at ease
with my emotions), 11 (My emotions make me uncom- 6 Because of limited space, communality estimates (i.e., the squared mul-
fortable), and 18 (When Im upset, I become scared and tiple correlation of each variable with the other variables in the analysis)
fearful of those feelings) had loadings of less than .40 on for the 36 items included in the final factor analysis could not be pub-
all factors, and item 36 (When Im upset, I know there lished. These data are available upon request from the authors.
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512
Assessment of Difficulties in Emotion Regulation
Table II. Eigenvalues and Percentage of Variance Accounted for by the Six Factors
in the Final Factor Analysis (N = 357)
Initial Extraction sums
eigenvalues of squared loadings
Factor Total % Variance Total
1 11.105 30.846 10.693
2 3.852 10.701 3.362
3 2.939 8.165 2.532
4 1.953 5.424 1.623
5 1.562 4.340 1.105
6 1.130 3.139 0.731
of variance accounted for by the six factors initially and
upon extraction).
The six factors comprising the DERS are highly inter-
pretable and reflect the multifaceted definition of emotion
regulation on which the scale was based (see Table III).
Factor 1 can be labeled Nonacceptance of Emotional Re-
sponses (NONACCEPTANCE). It is composed of items
reflecting a tendency to have negative secondary emo-
tional responses to ones negative emotions, or nonac-
cepting reactions to ones distress. Factor 2 can be la-
beled Difficulties Engaging in Goal-Directed Behavior
(GOALS). It is composed of items reflecting difficulties
concentrating and accomplishing tasks when experiencing
negative emotions. Factor 3 can be labeled Impulse Con-
trol Difficulties (IMPULSE) and is composed primarily of
items reflecting difficulties remaining in control of ones
behavior when experiencing negative emotions. Factor 4
can be labeled Lack of Emotional Awareness (AWARE-
NESS) and consists of items reflecting the tendency to
attend to and acknowledge emotions. When these items
are appropriately reverse-scored, this factor reflects an
inattention to, and lack of awareness of, emotional re-
sponses. Factor 5 can be labeled Limited Access to Emo-
tion Regulation Strategies (STRATEGIES) and consists
of items reflecting the belief that there is little that can
be done to regulate emotions effectively, once an individ-
ual is upset. Factor 6 can be labeled Lack of Emotional
Clarity (CLARITY) and is composed of items reflect-
ing the extent to which individuals know (and are clear
about) the emotions they are experiencing. As expected,
the six factors were correlated with one another (see
Table IV).
Results suggest that the four proposed dimensions
of emotion regulation on which the DERS is based may
be more accurately conceptualized as six distinct (albeit
related) dimensions. Although the NONACCEPTANCE
and STRATEGIES factors mapped directly onto the pro-
posed dimensions of emotional acceptance and access
November 22, 2003 15:36 Style file version June 25th, 2002
47
Rotation sums of
squared loadings
% Variance Total
29.703 5.585
9.339 6.704
7.033 6.764
4.509 3.874
3.068 9.074
2.030 6.071
to effective emotion regulation strategies, respectively,
results suggest that the two other proposed dimensions
of emotion regulation are themselves multidimensional.
Items thought to reflect difficulties in the proposed dimen-
sion of awareness and understanding of emotions loaded
onto two separate factors, AWARENESS and CLARITY,
suggesting that there may be a difference between being
aware of emotional responses and having a clear under-
standing of the nature of these responses. Similarly, items
representing difficulties in the proposed dimension involv-
ing the ability to engage in goal-directed behavior and re-
frain from impulsive behavior when experiencing negative
emotions were split into two separate factors, IMPULSE
and GOALS, suggesting that there may be a difference be-
tween being able to inhibit undesired behavior and engage
in desired behavior.
Mean scores for women and men on the overall
DERS as well as each subscale are shown in Table V. Gen-
der differences were found for only the AWARENESS
subscale, on which men had significantly higher scores
than women (t = 3.5, p < .01; for all other scales, ts <
1.37, ps > .10), suggesting that men reported lower emo-
tional awareness than women.
Reliability
Internal Consistency. Cronbachs was calculated
to determine the internal consistency of the DERS items.
Results indicate that the DERS had high internal con-
sistency ( = .93). Item-total correlations ranged from
r = .16 to r = .69. Thirty-four of the items had item-
total correlations above r = .30. All of the DERS sub-
scales (computed from the 6 factors obtained in the fac-
tor analysis) also had adequate internal consistency, with
Cronbachs > .80 for each subscale. See Table VI for
additional information on the internal consistency of each
subscale.
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512
48
Table III. Items Composing the Six DERS Factors
Factor
1: Nonacceptance of Emotional Responses
(NONACCEPTANCE)
2: Difficulties Engaging in Goal-Directed
Behavior (GOALS)
3: Impulse Control Difficulties (IMPULSE)
4: Lack of Emotional Awareness
(AWARENESS)
5: Limited Access to Emotion Regulation
Strategies (STRATEGIES)
6: Lack of Emotional Clarity (CLARITY)
Note. (r) = reverse-scored item.
Validity
Construct Validity. To provide preliminary data on
the construct validity of the DERS, correlations between
the DERS scores (the overall score as well as the indi-
vidual subscale scores) and a commonly used measure
of emotion regulation, the NMR, were computed. Cor-
relations were also computed between the DERS scores
and measures of experiential avoidance and emotional
expressivity. Since the DERS is based on a conceptual-
ization of emotion regulation that distinguishes adaptive
emotion regulation from emotional avoidance and
expressive control, evidence for the construct validity of
this measure of emotion dysregulation would be provided
by positive correlations with experiential avoidance and
November 22, 2003 15:36 Style file version June 25th, 2002
Gratz and Roemer
Item
29) When Im upset, I feel guilty for feeling that way.
25) When Im upset, I feel ashamed with myself for feeling that way.
15) When Im upset, I become embarrassed for feeling that way.
14) When Im upset, I become angry with myself for feeling that way.
33) When Im upset, I become irritated with myself for feeling that way.
27) When Im upset, I feel like I am weak.
30) When Im upset, I have difficulty concentrating.
22) When Im upset, I have difficulty focusing on other things.
16) When Im upset, I have difficulty getting work done.
38) When Im upset, I have difficulty thinking about anything else.
24) When Im upset, I can still get things done. (r)
37) When Im upset, I lose control over my behaviors.
31) When Im upset, I have difficulty controlling my behaviors.
17) When Im upset, I become out of control.
23) When Im upset, I feel out of control.
4) I experience my emotions as overwhelming and out of control.
28) When Im upset, I feel like I can remain in control of my behaviors. (r)
7) I am attentive to my feelings. (r)
3) I pay attention to how I feel. (r)
12) When Im upset, I acknowledge my emotions. (r)
21) When Im upset, I believe that my feelings are valid and important. (r)
9) I care about what I am feeling. (r)
39) When Im upset, I take time to figure out what Im really feeling. (r)
20) When Im upset, I believe that Ill end up feeling very depressed.
19) When Im upset, I believe that I will remain that way for a long time.
35) When Im upset, I believe that wallowing in it is all I can do.
40) When Im upset, it takes me a long time to feel better.
32) When Im upset, I believe that there is nothing I can do to make myself feel better.
26) When Im upset, I know that I can find a way to eventually feel better. (r)
41) When Im upset, my emotions feel overwhelming.
34) When Im upset, I start to feel very bad about myself.
6) I have difficulty making sense out of my feelings.
5) I have no idea how I am feeling.
10) I am confused about how I feel.
8) I know exactly how I am feeling. (r)
1) I am clear about my feelings. (r)
negative (rather than positive) correlations with emotional
expressivity.
As shown in Table VII, all correlations between the
overall DERS score and the constructs of interest were in
the expected directions and statistically significant. Each
of the DERS subscales was also significantly correlated
(in the expected direction) with the NMR and the mea-
sure of experiential avoidance; however, only three of the
subscales were significantly correlated with the measure
of emotional expressivity. Supporting the validity of the
factor solution, the DERS subscales showed a differen-
tial pattern of association with the constructs of inter-
est. Consistent with the expectation that the STRATE-
GIES subscale would capture difficulties in the aspect
of emotion regulation most commonly assessed in the
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512 November 22, 2003 15:36 Style file version June 25th, 2002

Assessment of Difficulties in Emotion Regulation 49

Table IV. Correlations Among Factors, Computed as Unweighted Sums other subscales (all ts > 4.28, ps < 001.), and the
of the Items Composing Each Scale (N = 357) CLARITY subscale was significantly more highly corre-
lated with the EES than were the NONACCEPTANCE,
Factor 1 2 3 4 5 6
GOALS, IMPULSE, and STRATEGIES subscales (all
1 ts > 2.02, ps < .05).
2 .33 To explore whether the DERS adds to the literature
3 .39 .50
on the assessment of difficulties in emotion regulation by
4 .14 .08 .22
5 .63 .62 .61 .16 accounting for additional variance in constructs of interest
6 .44 .32 .39 .46 .49 above and beyond that accounted for by an extant mea-
sure of emotion regulation, partial correlations between
Note. Factor 1 = Nonacceptance of Emotional Responses (NONACCEP- the DERS overall and subscale scores and the constructs
TANCE); Factor 2 = Difficulties Engaging in Goal-Directed Behavior
of interest were computed, controlling for the NMR. As
(GOALS); Factor 3 = Impulse Control Difficulties (IMPULSE); Factor
4 = Lack of Emotional Awareness (AWARENESS); Factor 5 = Limited shown in Table VII, all of the DERS variables accounted
Access to Emotion Regulation Strategies (STRATEGIES); Factor 6 = for a significant amount of additional variance in expe-
Lack of Emotional Clarity (CLARITY). riential avoidance, above and beyond that accounted for
p < .01.
by the NMR. In addition, the AWARENESS, CLARITY,
GOALS, and STRATEGIES subscales accounted for a
significant amount of additional variance in emotional ex-
literature, the correlation between the STRATEGIES sub- pressivity (although for the latter two subscales, the mag-
scale and the extant measure of emotion regulation (the nitude of the relationship was quite modest and the direc-
NMR) was significantly higher than the correlations be- tion was reversed).
tween the NMR and any of the other DERS subscales Predictive Validity. In order to begin to explore the
(all ts > 4.78, ps < .001; see Cohen & Cohen, 1983, predictive validity of the DERS (i.e., the extent to which
for the test of the significance of the difference between it is associated with clinically relevant behavioral out-
dependent r s). In addition, the correlation between the comes), correlations between the DERS scores and two
GOALS subscale and the NMR was significantly higher clinically important behavioral outcomes thought to be
than the correlations between the NMR and the NONAC- associated with emotion dysregulation (frequency of de-
CEPTANCE, AWARENESS, and CLARITY subscales liberate self-harm and frequency of intimate partner abuse)
(all ts > 2.18, ps < .05). The STRATEGIES subscale was were computed. Given gender differences in the severity
also significantly more highly associated with the measure and proposed function of partner abuse (see Saunders,
of experiential avoidance (the AAQ) than were the other 1986; Stets & Straus, 1990), as well as increasing evi-
DERS subscales (all ts > 3.15, ps < .01). As for the pat- dence of gender differences in the risk factors for self-
tern of associations between the DERS subscales and the harm (see Gratz, 2003; Gratz, Conrad, & Roemer, 2002),
measure of emotional expressivity (the EES), only three the relationships between difficulties in emotion regula-
of the subscales (AWARENESS, CLARITY, and NON- tion and these outcome variables were examined sepa-
ACCEPTANCE) were significantly correlated with the rately for women and men. A history of at least one inci-
EES. The AWARENESS subscale was significantly more dent of self-harm was reported by 35% of the women and
highly correlated with EES scores than were any of the 44% of the men (with more than 10 incidents reported
by 15% of the women and 14% of the men). A history
of at least one act of intimate partner abuse was reported
Table V. Means and Standard Deviations for DERS Scales Among by 24% of the women and 17% of the men (with 12%
Women (n = 260) and Men (n = 97) of the women and 8% of the men reporting more than
three acts of abuse in the past). These rates are consistent
Women Men
with past studies on these behaviors among other sam-
Scale Mean SD Mean SD ples from this urban university (see Gratz et al., 2002;
DERS Overall 77.99 20.72 80.66 18.79
Lisak, Hopper, & Song, 1996). Rates of self-harm and
NONACCEPTANCE 11.65 4.72 11.55 4.20 partner abuse did not differ significantly across gender
GOALS 14.41 4.95 14.34 5.16 ( 2 = 2.62 for self-harm and 2 = 2.10 for partner abuse;
IMPULSE 10.82 4.41 11.55 4.59 ps > .10).
AWARENESS 14.34 4.60 16.26 4.61 Before conducting analyses, logarithms were used to
STRATEGIES 16.16 6.19 16.23 6.26
CLARITY 10.61 3.80 10.74 3.67
transform the DSHI and PH frequency scores, as the raw
frequency scores of both measures were highly positively
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512 November 22, 2003 15:36 Style file version June 25th, 2002

50 Gratz and Roemer

Table VI. Internal Consistency Reliability Analyses for DERS Subscales (N = 357)

Range of Range of Mean

No. of Cronbachs item-total interitem interitem
Subscale items alpha correlations correlations correlation

NONACCEPTANCE 6 .85 .52.71 .33.67 .50

GOALS 5 .89 .59.81 .47.75 .62
IMPULSE 6 .86 .45.76 .31.73 .52
AWARENESS 6 .80 .49.67 .29.64 .40
STRATEGIES 8 .88 .46.75 .27.69 .47
CLARITY 5 .84 .56.71 .39.67 .51

skewed (as would be expected when assessing these con-
structs). Following transformations, both variables ap-
proximated normal distributions among female partici-
pants. Although the transformed DSHI scores approx-
imated a normal distribution among male participants,
their PH scores remained positively skewed and kur-
totic (skewness = 2.73, kurtosis = 7.63). However, Kline
(1998) reviewed Monte Carlo studies that indicate that
nonnormality is not problematic unless skewness >3 and
kurtosis >10; therefore, analyses were conducted using
these transformed variables for women and men.
As shown in Table VIII, the correlations between the
overall DERS score and self-harm were significant (and
in the expected direction) among both women and men,
and the correlation between the overall DERS score and
intimate partner abuse (also in the expected direction) was
significant among men. The overall DERS score was sig-
nificantly more highly correlated with frequency of partner
abuse among men than among women (z = 2.26, p < .05;
see Cohen & Cohen, 1983, for a test of the significance of
the difference between independent r s), consistent with
the emphasis in the literature on the emotion regulatory
function of partner abuse among men in particular.
Providing support for the validity of the factor so-
lution, the DERS subscales showed a differential pattern
of association with self-harm and intimate partner abuse,
within and across gender. Frequency of self-harm was sig-
nificantly associated with the NONACCEPTANCE and
IMPULSE subscales among men, but not among women
(for whom correlations with the other four subscales were
significant). However, none of the correlations between
frequency of self-harm behavior and the DERS subscales
differed significantly from one another among women or
men, and none differed significantly across gender.
As for the relationships between different aspects of
emotion regulation and intimate partner abuse, frequency
of partner abuse was significantly associated with only the
IMPULSE subscale among women, and with the GOALS,

Table VII. Correlations and Partial Correlations Among DERS Scales and Constructs of Interest (N = 357)

Scale NMR (N = 348) Experiential avoidance Emotional expressivity

DERS Overall .69 .60 .23

Controlling for NMRa .33 .09
NONACCEPTANCE .42 .39 .14
Controlling for NMRa .18 .05
GOALS .53 .44 .04
Controlling for NMRa .19 .11
IMPULSE .46 .41 .04
Controlling for NMRa .18 .09
AWARENESS .34 .32 .46
Controlling for NMRa .14 .42
STRATEGIES .69 .56 .10
Controlling for NMRa .28 .11
CLARITY .39 .38 .25
Controlling for NMRa .20 .18

Note. NMR = Generalized Expectancy for Negative Mood Regulation Scale.

aN = 348.
p < .05. p < .01.
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512 November 22, 2003 15:36 Style file version June 25th, 2002

Assessment of Difficulties in Emotion Regulation 51

Table VIII. Correlations and Partial Correlations Among DERS Scales and Clinical Outcome Variables of Frequency of
Self-Harm and Intimate Partner Abuse Among Women and Men

Self-harm Intimate partner abuse

Scale Women (N = 260) Men (N = 97) Women (N = 256) Men (N = 96)

DERS Overall .20 .26 .08 .34

Controlling for NMR .02a .16 .11b .25
NONACCEPTANCE .11 .28 .04 .18
Controlling for NMR .00a .21 .05b .09
GOALS .13 .13 .03 .37
Controlling for NMR .01a .02 .04b .30
IMPULSE .11 .24 .19 .29
Controlling for NMR .01a .16 .22b .20
AWARENESS .13 .07 .02 .10
Controlling for NMR .02a .04 .02b .06
STRATEGIES .20 .16 .05 .24
Controlling for NMR .03a .02 .07b .10
CLARITY .16 .16 .06 .13
Controlling for NMR .05a .12 .06b .08

Note. NMR = Generalized Expectancy for Negative Mood Regulation Scale.

aN = 251.
b N = 247.
p < .10. p < .05. p < .01.

IMPULSE, and STRATEGIES subscales among men. For
women, the correlation between the IMPULSE subscale
and frequency of partner abuse was significantly higher
than the correlations between the NONACCEPTANCE,
GOALS, AWARENESS, and STRATEGIES subscales
and frequency of partner abuse (ts > 2.16, ps < .05).
Among men, the correlation between the GOALS sub-
scale and frequency of partner abuse was significantly
higher than the correlations between the AWARENESS
and CLARITY subscales and frequency of partner abuse
(ts > 2.00, ps < .05). The GOALS subscale also was
significantly more highly correlated with frequency of
partner abuse among men than among women (z = 2.96,
p < .01). These findings support the proposed associa-
tion between difficulties in emotion regulation and part-
ner abuse among men, and suggest that for women this
relationship may exist only for the specific dimension of
difficulties controlling ones behavior when distressed.
To explore whether the DERS explains additional
variance in these clinical outcomes above and beyond
that accounted for by an extant measure of emotion reg-
ulation, partial correlations between the DERS overall
and subscale scores and these behavioral outcome mea-
sures were computed, controlling for the NMR. As shown
in Table VIII, there is some evidence that the DERS
accounts for unique variance in clinically relevant be-
haviors, when controlling for that accounted for by the
NMR. Specifically, the NONACCEPTANCE subscale ac-
counted for a significant amount of additional variance in
frequency of self-harm among men, whereas the GOALS
subscale accounted for a significant amount of additional
variance in frequency of partner abuse among men. In
addition, the IMPULSE subscale explained a significant
amount of additional variance in frequency of partner
abuse among women. When controlling for the NMR, the
overall DERS score explained a significant amount of the
variance in frequency of partner abuse among men.
STUDY 2
Method
Participants and Procedure
In order to assess the testretest reliability of the
DERS, a second sample was obtained. As part of a larger,
unrelated study, 194 participants were recruited from ta-
bles located in public areas on the University of Mas-
sachusetts Boston campus to complete a questionnaire
packet that included the DERS (in addition to other mea-
sures for the unrelated study). Of these participants, 21
agreed to complete the DERS again 48 weeks later, in ex-
change for a small monetary compensation. This subset of
participants ranged in age from 18 to 48 years, with a mean
age of 25.95 (SD = 8.94). Sixty-two percent of these par-
ticipants were female. In regard to the racial/ethnic back-
ground of these participants, 67% were White, 24% were
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512 November 22, 2003
52
Black/African American, 5% were Asian/Pacific Islander,
and 5% were of an unspecified racial/ethnic background.
This sample completed only the DERS, and not the other
measures described earlier.
Results
For the 21 participants in the testretest sample, in-
traclass correlation coefficients were computed for scores
on the first and second administration of the DERS in or-
der to determine the testretest reliability of the DERS
scores. Results indicate that the overall DERS score had
good testretest reliability over a period ranging from 4
to 8 weeks ( I = .88, p < .01). The testretest reliabil-
ity of the DERS subscales was adequate ( I s = .69 for
NONACCEPTANCE, .69 for GOALS, .57 for IMPULSE,
.68 for AWARENESS, .89 for STRATEGIES, and .80 for
CLARITY; all ps < .01).
GENERAL DISCUSSION
The results of these studies contribute to and ex-
pand upon literature on the conceptualization of emo-
tion regulation and provide preliminary support for the
DERS as a measure of difficulties in emotion regula-
tion. Results suggest the presence of six separate (albeit
related) dimensions of emotion regulation wherein dif-
ficulties may occur, including (a) lack of awareness of
emotional responses, (b) lack of clarity of emotional re-
sponses, (c) nonacceptance of emotional responses, (d)
limited access to emotion regulation strategies perceived
as effective, (e) difficulties controlling impulses when ex-
periencing negative emotions, and (f) difficulties engag-
ing in goal-directed behaviors when experiencing negative
emotions. These findings suggest the importance of dis-
tinguishing between the awareness and understanding of
emotional responses, as well as between the ability to act
in desired ways and refrain from acting in undesired ways
when experiencing negative emotions.
Findings provide empirical support for a multidimen-
sional conceptualization of emotion regulation. Correla-
tions of the DERS variables with several clinically relevant
constructs revealed a differential pattern of associations
amongst the different DERS subscales, with certain sub-
scales showing stronger relationships with particular con-
structs and/or behavioral outcomes than others. Significant
associations between DERS subscales assessing deficits
in awareness, understanding, and acceptance of emotions
and the clinically relevant constructs of emotional avoid-
ance and expressivity support the comprehensive concep-
tualization of emotion regulation on which the DERS is
based. Furthermore, the significant relationships between
15:36 Style file version June 25th, 2002
Gratz and Roemer
these subscales and the behavioral outcome of self-harm
among both women and men (specifically, the AWARE-
NESS and CLARITY subscales among women and the
NONACCEPTANCE subscale among men) suggest their
clinical relevance. Findings that the subscales focusing on
difficulties controlling behavior when emotion is present
(i.e., GOALS and IMPULSE), rather than on difficulties
controlling emotion, were significantly related to clini-
cal outcomes suggest the usefulness of considering the
impact of these particular dimensions within the clinical
realm.
Although the DERS and its subscales accounted for
additional variance in clinical constructs and behavioral
outcomes above and beyond that accounted for by the
NMR, many of the relationships between the DERS scales
and the behavioral outcomes did not remain statistically
significant after controlling for the NMR. This is likely
due to the rather high correlations between the NMR and
some of the DERS subscales, suggesting that despite the
NMRs theoretical emphasis on emotional avoidance and
inclusion of emotional control items, there is some overlap
between the DERS and NMR. Nevertheless, the fact that
some unique relationships remained between the DERS
and behavioral outcomes after controlling for the NMR
suggests that the NMR does not capture all clinically rel-
evant aspects of emotion regulation. It will be important
to continue to investigate the clinical utility of the DERS
in explaining a broad range of behavioral outcomes.
Despite preliminary findings that the DERS has high
internal consistency, good testretest reliability, and ade-
quate construct and predictive validity, the present stud-
ies have limitations that should be addressed in future
research. One limitation is that the testretest reliability
results are based on a small sample size (N = 21), thus
requiring replication with a larger sample. Another limi-
tation is the reliance on a few select self-report measures
of emotional responding to provide data on the construct
validity of the DERS. As a result, the relationships be-
tween the DERS and other aspects of emotional respond-
ing remain to be determined. Moreover, the reliance on
only self-reported emotional responding is limiting, as it
is likely that some individuals do not have full awareness
of their emotional responses, thereby reducing the extent
to which they can accurately report on those responses.
Despite the comprehensive nature of the DERS, omit-
ted constructs should be noted. DERS items focus primar-
ily on the regulation of negative emotional states (i.e.,
many items begin with the phrase When Im upset), as
difficulties in this domain are likely to have the most clin-
ical relevance. Future research may want to incorporate
items assessing difficulties with the regulation of positive
emotional states as well. Also, although results support
the clinical utility of the NONACCEPTANCE subscale, it
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512
Assessment of Difficulties in Emotion Regulation
is important to note that the items included in this subscale
assess only one aspect of the nonacceptance of emotional
responses, secondary emotional reactions to ones nega-
tive emotions. This form of emotional nonacceptance was
emphasized due to its clinical relevance and proposed as-
sociation with negative emotional and behavioral conse-
quences (see Greenberg & Paivio, 1997); however, future
research on scale development should include items as-
sessing other aspects of emotional nonacceptance.
The extent to which the results of this study can
be generalized to men or individuals from diverse eth-
nic/racial backgrounds remains an empirical question. Al-
though the sample comes from a diverse urban univer-
sity that draws heavily from the community and attracts a
large number of older, nontraditional college students, the
majority of the participants were White women. Given
evidence of the role of gender socialization and culture
in emotional appraisal and expression (see Kitayama &
Markus, 1997; Manstead, 1992), it is possible that the fac-
tor structure of emotion regulation and/or the relationships
of these factors to clinical constructs differs as a function
of gender or cultural background. It is also important to
note that the validity of the DERS for a clinical population
has yet to be determined (see Haynes, Richard, & Kubany,
1995).
Given the preliminary nature of this study, replica-
tion of the results with different samples, under differ-
ent circumstances (e.g., different assessment procedures),
and across different functions is necessary to ensure the
robustness and generalizability of the findings (Haynes
et al., 1995). Examination of the psychometric properties
and factor structure of the DERS among diverse nonclin-
ical populations, as well as relevant clinical populations
(such as individuals with borderline personality disorder;
see Linehan, 1993), is necessary. Future research on the
validity of the DERS should examine its relationships with
a broader range of self-report measures of emotional re-
sponding (such as the TMMS and other measures of emo-
tional awareness and understanding), as well as assess-
ments of emotional responding in other domains (e.g.,
physiological responding, the coding of facial expressiv-
ity). Further research is also needed to examine the dis-
criminant validity of the DERS, especially with respect
to other measures of general distress or psychopathology,
such as anxiety and depression.
Further research is needed on the predictive validity
of the DERS as well. The behavioral outcomes of self-
harm and intimate partner abuse were chosen to provide
preliminary data in this area in part because of the the-
oretical support for the role of emotion dysregulation in
both behaviors, and in part because of their interest to
these authors. There are many other clinically relevant
behaviors and constructs that are also thought to be asso-
November 22, 2003 15:36 Style file version June 25th, 2002
53
ciated with deficits in emotion regulation, including post-
traumatic stress disorder and generalized anxiety disorder,
and future research on the relationships between the DERS
subscales and these clinical outcomes is needed. Research
exploring the differential role of the various dimensions
of emotion dysregulation in clinical problems will be es-
pecially important, as it may suggest specific targets for
intervention.
ACKNOWLEDGMENTS
This research was supported in part by a grant to the
first author from the Office of Research and Sponsored
Programs at the University of Massachusetts Boston. The
authors wish to thank Matthew Tull, Amy Wagner, and
two anonymous reviewers for their helpful comments on
earlier versions of the measure and/or previous drafts of
this manuscript.
REFERENCES
Briere, J., & Gil, E. (1998). Self-mutilation in clinical and general pop-
ulation samples: Prevalence, correlates, and functions. American
Journal of Orthopsychiatry, 68, 609620.
Bushman, B. J., Baumeister, R. F., & Phillips, C. M. (2001). Do people
agree to improve their mood? Catharsis beliefs, affect regulation
opportunity, and aggressive responding. Journal of Personality and
Social Psychology, 81, 1732.
Catanzaro, S. J., & Mearns, J. (1990). Measuring generalized expectan-
cies for negative mood regulation: Initial scale development and
implications. Journal of Personality Assessment, 54, 546563.
Cloitre, M. (1998). Sexual revictimization: Risk factors and prevention.
In V. M. Follette, J. I., Ruzek, & F. R. Abueg (Eds.), Cognitive
behavioral therapies for trauma (pp. 278304). New York: Guilford
Press.
Cohen, J., & Cohen, P. (1983). Applied multiple regression/correlation
analysis for the behavioral sciences (2nd ed.). Hillsdale, NJ:
Erlbaum.
Cole, P. M., Michel, M. K., & Teti, L. O. (1994). The development of
emotion regulation and dysregulation: A clinical perspective. In N.
A. Fox (Ed.), The development of emotion regulation: Biological
and behavioral considerations. Monographs of the Society for Re-
search in Child Development, 59 (Serial No. 240), 73100.
Cortez, V. L., & Bugental, D. B. (1994). Childrens visual avoidance of
threat: A strategy associated with low social control. Merrill-Palmer
Quarterly, 40, 8297.
Eisenberg, N., Cumberland, A., & Spinrad, T. L. (1998). Parental social-
ization of emotion. Psychological Inquiry, 9, 241273.
Flett, G. L., Blankstein, K. R., & Obertynski, M. (1996). Affect inten-
sity, coping styles, mood regulation expectancies, and depressive
symptoms. Personality and Individual Differences, 20, 221228.
Floyd, F. J., & Widaman, K. F. (1995). Factor analysis in the development
and refinement of clinical assessment instruments. Psychological
Assessment, 7, 286299.
Garner, P. W., & Spears, F. M. (2000). Emotion regulation in low-income
preschoolers. Social Development, 9, 246264.
Gratz, K. L. (2001). Measurement of deliberate self-harm: Prelimi-
nary data on the Deliberate Self-Harm Inventory. Journal of Psy-
chopathology and Behavioral Assessment, 23, 253263.
Gratz, K. L. (2003). Risk factors for and functions of deliberate self-
harm: An empirical and conceptual review. Clinical Psychology:
Science and Practice, 10, 192205.
P1: GXB
Journal of Psychopathology and Behavioral Assessment (JOBA) PP1052-joba-476512
54
Gratz, K. L., Conrad, S. D., & Roemer, L. (2002). Risk factors for de-
liberate self-harm among college students. American Journal of
Orthopsychiatry, 72, 128140.
Greenberg, L. S., & Paivio, S. C. (1997). Working with emotions in
psychotherapy. New York: The Guilford Press.
Gross, J. J., & Levenson, R. W. (1997). Hiding feelings: The acute effects
of inhibiting negative and positive emotion. Journal of Abnormal
Psychology, 106, 95103.
Gross, J. J., & Munoz, R. F. (1995). Emotion regulation and mental
health. Clinical Psychology: Science and Practice, 2, 151164.
Hayes, S. C., Strosahl, K. D., Wilson, K. G., Bissett, R. T., Pistorello, J.,
Toarmino, D. et al. (in press). Measuring experimental avoidance:
A preliminary test of a working model. The Psychological Record.
Hayes, S. C., Strosahl, K. D., & Wilson, K. G. (1999). The ACT model of
psychopathology and human suffering. In Acceptance and commit-
ment therapy: An experiential approach to behavior change (pp. 49
80). New York: Guilford Press.
Hayes, S. C., Wilson, K. G., Gifford, E. V., Follette, V. M., &
Strosahl, K. (1996). Experiential avoidance and behavioral disor-
ders: A functional dimensional approach to diagnosis and treat-
ment. Journal of Consulting and Clinical Psychology, 64, 1152
1168.
Haynes, S. N., Richard, D. C. S., & Kubany, E. S. (1995). Content validity
in psychological assessment: A functional approach to concepts and
methods. Psychological Assessment, 7, 238247.
Jakupcak, M., Lisak, D., & Roemer, L. (2002). The role of masculine
ideology and masculine gender role stress in mens perpetration of
aggression and violence in relationships. Psychology of Men and
Masculinity, 3, 97106.
Kendall, M. G., & Stuart, A. (1958). The advanced theory of statistics.
New York: Hafner.
Kitayama, S., & Markus, H. R. (1997). Emotion and culture: Empirical
studies of mutual influence. Washington, DC: American Psycho-
logical Association.
Kline, R. B. (1998). Principles and practices of structural equation mod-
eling. New York: Guilford Press.
Kopp, C. B. (1989). Regulation of distress and negative emotions: A
developmental view. Developmental Psychology, 25, 343354.
Kring, A. M., Smith, D. A., & Neale, J. M. (1994). Individual differences
in dispositional expressiveness: The development and validation of
the Emotional Expressivity Scale. Journal of Personality and Social
Psychology, 66, 934949.
Linehan, M. M. (1993) Cognitivebehavioral treatment of borderline
personality disorder. New York: The Guilford Press.
Lisak, D., Conklin, A., Hopper, J., Miller, P., Altschuler , L. & Smith,
B. (2000). The Abuse-Perpetration Inventory: Development of an
assessment instrument for research on the cycle of violence. Family
Violence and Sexual Assault Bulletin, 2130.
November 22, 2003 15:36 Style file version June 25th, 2002
Gratz and Roemer
Lisak, D., Hopper, J., & Song, P. (1996). Factors in the cycle of violence:
Gender rigidity and emotional constriction. Journal of Traumatic
Stress, 9, 735757.
Manstead, A. S. R. (1992). Gender differences in emotion. In A. Gale
and M. W. Eysenck (Eds.), Handbook of individual differences:
Biological Perspectives (pp. 355387). Chichester, England: Wiley.
Melnick, S. M., & Hinshaw, S. P. (2000). Emotion regulation and parent-
ing in AD/HD and comparison boys: Linkages with social behav-
iors and peer preference. Journal of Abnormal Child Psychology,
28, 7386.
Mennin, D. S., Heimberg, R. G., Turk, C. L., & Fresco, D. M. (2002). Ap-
plying an emotion regulation framework to integrative approaches
to generalized anxiety disorder. Clinical Psychology: Science and
Practice, 9, 8590.
Notarius, C. I., & Levenson, R. W. (1979). Expressive tendencies and
physiological response to stress. Journal of Personality and Social
Psychology, 37, 12041210.
Paivio, S. C., & Greenberg, L. S. (1998). Experiential theory of emo-
tion applied to anxiety and depression. In W. F. Flack Jr. & J. D.
Laird (Eds.), Emotions in psychopathology: Theory and research
(pp. 229242). New York: Oxford University Press.
Saarni, C. (1999). The development of emotional competence. New York:
Guilford Press.
Salovey, P., Mayer, J. D., Goldman, S. L., Turvey, C., & Palfai, T. P.
(1995). Emotional attention, clarity and repair: Exploring emotional
intelligence using the trait meta-mood scale. In J. W. Pennebaker
(Ed.), Emotion, disclosure, and health (pp. 125154). Washington,
DC: APA Books.
Saunders, D. G. (1986). When battered women use violence: Husband
abuse or self-defense? Violence and Victims, 1, 4760.
Stets, J. & Straus, M. A. (1990). Gender differences in reporting mar-
ital violence and its medical and psychological consequences. In
M. A. Straus, & R. J. Gelles (Eds.), Physical violence in Amer-
ican families: Risk factors and adaptations in American families
(pp. 151166). New Brunswick, NJ: Transaction Books.
Stewart, S. H., Zvolensky, M. J., & Eifert, G. H. (2002). The relations of
anxiety sensitivity, experiential avoidance, and alexithymic coping
to young adults motivations for drinking. Behavior Modification,
26, 274296.
Thompson, R. A. (1994). Emotion regulation: A theme in search of def-
inition. In N. A. Fox (Ed.), The development of emotion regulation:
Biological and behavioral considerations. Monographs of the Soci-
ety for Research in Child Development, 59 (Serial No. 240), 2552.
Thompson, R. A., & Calkins, S. D. (1996). The double-edged sword:
Emotional regulation for children at risk. Development and Psy-
chopathology, 8, 163182.
Zeman, J. & Garber, J. (1996). Display rules for anger, sadness, and pain:
It depends on who is watching. Child Development, 67, 957973.
J Psychopathol Behav Assess
DOI 10.1007/s10862-008-9102-4

ERRATUM

Multidimensional Assessment of Emotion Regulation

and Dysregulation: Development, Factor Structure,
and Initial Validation of the Difficulties in Emotion
Regulation Scale
Kim L. Gratz & Lizabeth Roemer

# Springer Science + Business Media, LLC 2008

Erratum to: J Psychopathol Behav Assess (2004) 26:41-54
DOI 10.1023/B:JOBA.0000007455.08539.94
In stating that the Generalized Expectancy for Negative
Mood Regulation Scale (NMR; Catanzaro & Mearns, 1990)
was used as a template to help structure the format (but not
the content) of the Difficulties in Emotion Regulation Scale
(DERS), we failed to specify that one of the DERS items
(# 35, When Im upset, I believe that wallowing in it is all
I can do.) is identical in wording to an NMR item, and
another (#20, When Im upset, I believe that Ill end up
feeling [very] depressed.) is almost identical in wording,
while three other DERS items (# 19, 32, and 34) are similar
in content to NMR items. Thus, 5 of the 36 items in the
DERS are content-related to the NMR. Further, three other
DERS items (# 12, 21, and 39) were adapted from items in
the Emotional Approach Coping Scale (Stanton, Kirk,
Cameron, & Danoff-Burg, 2000) and four others overlap
with items from the Trait Meta-Mood Scale (Salovey,
Mayer, Goldman, Turvey, & Palfai, 1995). Finally, to
clarify the statement that many DERS items begin with
When Im upset, similar to the NMR (p. 44), it should be
noted that 27 of the 36 items begin with this phrase (just as
all of the NMR items do).

The online version of the original article can be found at http://dx.doi.

org/10.1023/B:JOBA.0000007455.08539.94.
K. L. Gratz : L. Roemer
Department of Psychology, University of Massachusetts Boston,
Boston, MA, USA

Present address:
K. L. Gratz (*)
Department of Psychiatry and Human Behavior,
University of Mississippi Medical Center,
2500 North State Street,
Jackson, MS 39216, USA
e-mail: KLGratz@aol.com